On The Evaluating Approaches for The Safety about Short-Term Loans of Bank

• Main Authors: CHANG,PO-SHU, 張柏樹
• Other Authors: 吳祥華
• Format: Others
• Language: zh-TW
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/48417504428266956404

碩士 === 國立臺北大學 === 統計學系 === 93 === According to the financial statistics monthly bulletin of the National Central Bank, our country has established 47 banks by March 2005 and the branches are up to 3202. Therefore, the competition of banking is vigorous and it is not easy to keep going on. According to the domestic bank ecosystem, each bank’s short-term loan to the enterprise is still the most important business. So how to control business credit risk is the important key to the constructive risk management. Banking needs to establish an effective credit ratings of enterprise and early warning mode of financial crisis to avoid the contract-breaking risk of the lending enterprise. The general credit ratings institutions often waste a lot of manpower and time cost to achieve the reliable information of credit ratings. They cannot satisfy the trade and urgency of lending, and they need to finish the security evaluation in the shortest time and the less lending manpower. So this research has huge scope of development. On the selection of the financial indicators, we can consult the ten financial indicators chosen by TCRI credit ratings. In addition, we aim at the security of banking short-term lending and use the relative financial indicators of corporate liquidity, cash flow and in-debt phenomena to be the reference variable of establishing mode. In way of doing, we bring into 39 financial indicators as the mode variable to sieve. As for the sample information, we choose the corporations which has had financial crisis in domestic public publishing ones after 1998 to match TCRI ratings1 level to 4 level financial normal corporation. Besides, in the process of research we pick out the sample of untrue financial contents and for financial lag, we cannot response to the information immediately. Finally, we choose the most representative indicator and use statistic mode—mainly Logit Regression Model, Discriminate Analysis Model, …etc to compare and analyze, and establish the whole ratings institution. On the evaluation of mode ratings results, compared with the discrimination of discriminate analysis, we found that which of Logit Regression Analysis is better. The evaluation accuracy of the earlier sample is up to 99.2% and 97.4% separately, and which of the sample to be retained is up to 97.5% and 95% separately. KEY WORDS：TCRI、Credit ratings、Contract-breaking risk